That's good new Gary! I suspect you will be pleasantly surprised at the results. Well, I hope so. You may need some tweaking from the sims, but everything presented should make for a really nice speaker in the end to say the least. I only wish I could get to hear them. I presume they will be at an audiofest sometime/somewhere? RMAF?
Dan
Dan
Thanks Dan!
After settling on the values for the Zobel, I added the low pass filter using calculated values for a 2nd order Bessel at 700 Hz. Of course, the lift from 1.5 kHz to 4 kHz was still there with all its ripples, but the whole thing was rotated clockwise. The next step was to knock down the first big bump with a narrow notch filter. I experimented with various notch tunings, widths and depths. Best results were achieved with a filter centered at 1,720 Hz. IR improved as well.
There was still a problem, though: I wasn't reaching -6 dB until about 900 Hz. Experimentation with added capacitance and inductance eventually resulted in a curve that hit the target rolloff and had the desired shape. The filter was now electrically tuned for -6 dB at 420 Hz! Thank goodness for Arta.
Best of all, the IR and CSD plots looked better than ever. The 2.5 ms wrinkle, while still present, is now a tiny fraction of its former self. I am very pleased with where it ended up. I think it's as good as I'm going to get with a 2nd order filter on a 15" woofer.
Here's the FR curve:
Here's the impulse response:
And here's the CSD:
Finally, here is the crossover schematic:
Gary Dahl
After settling on the values for the Zobel, I added the low pass filter using calculated values for a 2nd order Bessel at 700 Hz. Of course, the lift from 1.5 kHz to 4 kHz was still there with all its ripples, but the whole thing was rotated clockwise. The next step was to knock down the first big bump with a narrow notch filter. I experimented with various notch tunings, widths and depths. Best results were achieved with a filter centered at 1,720 Hz. IR improved as well.
There was still a problem, though: I wasn't reaching -6 dB until about 900 Hz. Experimentation with added capacitance and inductance eventually resulted in a curve that hit the target rolloff and had the desired shape. The filter was now electrically tuned for -6 dB at 420 Hz! Thank goodness for Arta.
Best of all, the IR and CSD plots looked better than ever. The 2.5 ms wrinkle, while still present, is now a tiny fraction of its former self. I am very pleased with where it ended up. I think it's as good as I'm going to get with a 2nd order filter on a 15" woofer.
Here's the FR curve:
Here's the impulse response:
And here's the CSD:
Finally, here is the crossover schematic:
Gary Dahl
Sorry about the crude quality of my hand-drawn sketch for the crossover.
This has been a fun collaboration with Gary. I e-mailed him the sketch, and suggested that he start by designing the Zobel correctors for both drivers by direct measurement of the driver's impedance curve and hand-trimming the physical RC's to flatten the curve as much as possible in the Z-minimum region and at higher frequencies. This typically gives a rather different result than computer simulation, partly because real-world woofer inductance is not actually a simple inductance, but a more complex term.
Once that was done, I suggested starting with the lowpass filter and first seeing if a basic 2nd-order lowpass filter had the right shape. It didn't. The bump at 1.5 ~ 1.8 kHz fell in an awkward part of the rolloff curve, and would have adversely affected the inter-driver phase angle in a rather sensitive part of the audio spectrum. So I suggested going through a range of possible notch filters using a shunt topology - and Gary zeroed in on a good candidate rather quickly, cross-checking in the time domain to make sure the time response was really improving as well. Tuning notch filters with real-world drivers requires frequent comparisons of frequency and time-domain performance; when both improve together, the filter is working correctly.
The next step is a complementary highpass filter for the AH425 and Radian 745P. The frequency-response, impulse-response, and impedance measurements are promising, so the highpass part of the filter should be straightforward, and the heavy impedance damping of the shunt resistor across the "primary" of the autoformer acts to smooth out the impedance curve of the driver/horn combination. Since the horn is intended for free-air use, sitting above the bass cabinet on a wooden cradle, we are free to move it back and forth to optimize the phase relationship between the bass driver and the horn.
It also helps that the bass driver now has a smooth and well-defined lowpass characteristic, which greatly simplifies the integration between the two drivers. The goal is to get a phase-match between the two drivers of 10 degrees or better in the crossover region, and maintaining that phase-match over a reasonably wide listening angle.
Considering I'm in Colorado and Gary is in Washington State, and we're communicating by telephone and e-mail, it's gone rather quickly. I've joked with him he now knows more about crossover design than a number of high-name-recognition audiophile designers, who can't seem to make flat loudspeakers, and don't know how to tune notch filters.
Of course, the crowd here at diyAudio could do this kind of system integration in their sleep, but unfortunately that level of knowledge is far and few between in the audiophile high-end market (as a visit to the CES or RMAF will reveal).
This has been a fun collaboration with Gary. I e-mailed him the sketch, and suggested that he start by designing the Zobel correctors for both drivers by direct measurement of the driver's impedance curve and hand-trimming the physical RC's to flatten the curve as much as possible in the Z-minimum region and at higher frequencies. This typically gives a rather different result than computer simulation, partly because real-world woofer inductance is not actually a simple inductance, but a more complex term.
Once that was done, I suggested starting with the lowpass filter and first seeing if a basic 2nd-order lowpass filter had the right shape. It didn't. The bump at 1.5 ~ 1.8 kHz fell in an awkward part of the rolloff curve, and would have adversely affected the inter-driver phase angle in a rather sensitive part of the audio spectrum. So I suggested going through a range of possible notch filters using a shunt topology - and Gary zeroed in on a good candidate rather quickly, cross-checking in the time domain to make sure the time response was really improving as well. Tuning notch filters with real-world drivers requires frequent comparisons of frequency and time-domain performance; when both improve together, the filter is working correctly.
The next step is a complementary highpass filter for the AH425 and Radian 745P. The frequency-response, impulse-response, and impedance measurements are promising, so the highpass part of the filter should be straightforward, and the heavy impedance damping of the shunt resistor across the "primary" of the autoformer acts to smooth out the impedance curve of the driver/horn combination. Since the horn is intended for free-air use, sitting above the bass cabinet on a wooden cradle, we are free to move it back and forth to optimize the phase relationship between the bass driver and the horn.
It also helps that the bass driver now has a smooth and well-defined lowpass characteristic, which greatly simplifies the integration between the two drivers. The goal is to get a phase-match between the two drivers of 10 degrees or better in the crossover region, and maintaining that phase-match over a reasonably wide listening angle.
Considering I'm in Colorado and Gary is in Washington State, and we're communicating by telephone and e-mail, it's gone rather quickly. I've joked with him he now knows more about crossover design than a number of high-name-recognition audiophile designers, who can't seem to make flat loudspeakers, and don't know how to tune notch filters.
Of course, the crowd here at diyAudio could do this kind of system integration in their sleep, but unfortunately that level of knowledge is far and few between in the audiophile high-end market (as a visit to the CES or RMAF will reveal).
Last edited:
Lynn, I thought that crude sketch was was the best part. My sketch schematics look very sloppy by comparison! You even added shadowing/shading to speakers. I'd really enjoy subjective impressions of the "with tweeter" and "without tweeter" designs. To me it looks like throwing a super tweeter in the mix is sort of unnecessary overkill--if you can believe that(an audiophile saying something is overkill). I do understand why you'd want it and I am interested in your estimate of its audibility. My bet would be pretty small but worth it, cost no object, to make a statement type project.
Dan
Dan
Audibility, in terms of tonality, will be low to nonexistent. There's very little sensation of pitch above 7~10 kHz. Anyone's that's ever played with an old mono preamp has probably noticed that a 7~10 kHz rolloff, if done with a proper 2nd-order filter, is harder to hear than you might expect, give the great audibility of everything just an octave lower.
Mostly it sounds more shut-in, a bit less airy and "hifi" - in other words, the main perception is a difference in spaciousness. This isn't that surprising; the realm of short wavelengths is also the realm where the ridges and reflections of the outer ear, the pinna, are used to analyze directionality.
A notch at 7 kHz is very different-sounding than a notch at 3.5 or 1.75 kHz. At the lower frequencies, there are obvious tonal shifts, depending on the width and depth of the notch, and whether musical harmonics land on the notch frequency. At 7 kHz and above, a notch might produce an elevation artifact, raising the apparent height of the image source, since the notch interacts with the reflections and comb-filtering of the outer ear.
So the main purpose of a supertweeter (any tweeter working at 7 kHz or above is a supertweeter) is to improve the sensation of spaciousness - to "open up" the soundstage. This works hand-in-hand with the narrowing directivity of the T= 0.707 LeCleac'h horn and the 1.4" throat size of the large-format driver. Unlike a small-format horn which aims at reaching 15 kHz or above, this horn gives up the top octave in return for another octave at the bottom of the range. The diaphragm of a large-format driver is twice as large as a small-format driver, and that's the tradeoff.
You wouldn't expect a 15" bass driver to behave in the same way as a 7" midbass; they cover different frequency ranges, and have different power-handling characteristics. Similarly, the large and small-format drivers do different tasks.
I've built speakers with supertweeters before, and the crossover is a nuisance to measure, since the drivers are several wavelengths apart. This makes microphone positioning rather critical; move it up or down a few inches, and the response shows the typical notches of a speaker with tightly spaced vertical lobing. Intuitively, this seems like a bad thing, since you know the listening height is not that well controlled, and the listener might even tilt their head enough to create dissimilar responses at each ear.
But - in practice, it is remarkably difficult to hear. It really is, although very obvious in measurement. This is just like time distortion of square waves; it looks just dreadful on a scope, yet is not easy to hear. I've flipped the switch on all-pass filters and was hard-pressed to hear any difference - at most, a very subtle change in tonality, nothing like the clear change a 1/2 dB level difference in the midrange can make.
The supertweeter crossover, at least if it reasonably competent, is also more subtle than you might expect. The one thing to avoid is having the supertweeter louder than the main speaker; the shelved-up response is subjectively experienced as a peak at the frequency where the shelf begins. So if there's going to be an error, it should always be in the direction of having the supertweeter slightly down in level, and never higher. That does sound unnatural.
This isn't surprising; in nature, there are very few mechanisms that elevate treble over bass, while merely passing sound through 50 feet or more of air will dull the upper treble. A slight dullness merely sounds like the music is further away. Trim the supertweeter level, and you adjust the sense of "air" at the top - go too far, and the sound becomes unnatural and artificial-sounding.
One unusual aspect of ribbon tweeter's narrow vertical directionality is they don't fall off with distance as fast as direct-radiators. Although a ribbon and a dome tweeter might have equal levels at 1 meter, at 3 meters the ribbon pulls ahead by several dB because of less attenuation with distance.
In practice, you never want to equalize a speaker at 1 meter; the only speakers that are listened at that distance are nearfield monitors mounted on top of the meter bridge of a recording console. This system is never going to mounted on top of a console, nor will it be listened to at 1 meter. A 2.5 to 4 meter listening distance should be more typical.
If the directivity index of all the drivers were identical and invariant with frequency, sure, we could equalize at any distance and the speaker would behave just the same. In practice, though, the drivers have different directivity indexes, particularly at the edges of their working ranges. This means the final balancing needs to account for listening distance, or at the least, choose an average distance of say, 3 meters, and balance the speaker for that.
I might as well step in the big mud puddle of the directivity debate, now that I've gone this far. I believe in equalizing the speaker for the most likely listening positions, but do not agree with modern constant-directivity theory.
The reason is fairly simple; the first-arrival sound dominates the perception of spectral balance, while the first reflections come from wildly different emission angles, far outside the usual 60~90 degree "good window" of constant-directivity designs. We don't have to guess where the first reflections are coming from; we know what they are.
The first is the floor bounce; this will arrive 2.5~3 mSec after the direct sound. Most people will use nothing more than carpet (and foam or rubber underlayment) to absorb floor bounce. I've measure the spectra of the reflection off heavy carpet, and there is essentially no absorption below 7 kHz. So there floor bounce is there, whether we want it or not.
If the listener is at 48 inches (unlikely), the center of the speaker is also at 48 inches, and the ceiling is 8 feet, the ceiling reflection will arrive at the same time as the floor bounce. With a lower listening position and a higher ceiling, the ceiling reflection will arrive 1 mSec or more after the floor bounce.
Side walls and rear walls: if the speaker is jammed in a corner, the nearest sidewall reflection will be in the same range as the floor bounce - for obvious and some non-obvious reasons, this is not desirable. The same applies to the rear wall reflection - if this arrives at the same time as the floor bounce, or side wall reflection, this creates a noticeable coloration, since the reflections add up, and deepen the comb-filter nulls and peaks. In practice, moving the speaker out a moderate distance into the room spreads the reflections apart, leaving the floor bounce as the first reflection that arrives at the listening position. The others - ceiling, sidewall, and rear wall - follow several milliseconds later. Multiple bounces arrive much later, 10 mSec or more, unless the room is very small.
All four reflections have one thing in common: the emission angles for these first reflections are well outside the 60-90 degree controlled-directivity radiation angle. In fact, you can't hear the emission of the speaker 10 or 20 degrees away from where you're sitting; you have to physically move there to intercept the sound with your ears. When you sit in one location, you hear the sound that's coming right at you, plus the four reflections mentioned above, then an array of later reflections. All of the reflections take direct paths from the speaker to the reflection surface to you, and the great majority of emission angles for these reflections are far outside the 60-90 degree "good performance" window.
As far as nearly all these reflections are concerned, the emission angle is nearly random - not just from the front of the speaker, but the sides, top, bottom, and rear are equally weighted in terms of room reflections.
That is why I give greatest weight to the first-arrival sound - for my preference, maybe a weighting of 60~75% in terms of design priorities, which is quite different than the modern school of constant-directivity loudspeaker design. But these are the reasons why: you can't sit in five places at once, and domestic loudspeakers are not intended theater or PA use, where audience coverage is the prime directive. Sure, the listener in the living room might be well off-axis, and the dispersion should be smooth and even enough to cover all likely listening positions. But the energy going into the room reflections isn't coming from the preferred 60-90 emission angle; it's coming from all over the speaker, not just the forward radiation pattern.
So I give first priority to direct-arrival sound, and then consider the spectra of the total power of the speaker radiated into a sphere, since that is what is powering the reflections - they really do come from all directions at once, not just a preferred emission angle in the forward direction. With a multiway speaker with non-coincident drivers, the time profile of the speaker will be radically different for every reflection, since the emission angles are so radically different. The time profile that matters is the first arrival; all the rest will be thoroughly scrambled, unless you have a near-perfect single-point emitter (with no crossovers).
Mostly it sounds more shut-in, a bit less airy and "hifi" - in other words, the main perception is a difference in spaciousness. This isn't that surprising; the realm of short wavelengths is also the realm where the ridges and reflections of the outer ear, the pinna, are used to analyze directionality.
A notch at 7 kHz is very different-sounding than a notch at 3.5 or 1.75 kHz. At the lower frequencies, there are obvious tonal shifts, depending on the width and depth of the notch, and whether musical harmonics land on the notch frequency. At 7 kHz and above, a notch might produce an elevation artifact, raising the apparent height of the image source, since the notch interacts with the reflections and comb-filtering of the outer ear.
So the main purpose of a supertweeter (any tweeter working at 7 kHz or above is a supertweeter) is to improve the sensation of spaciousness - to "open up" the soundstage. This works hand-in-hand with the narrowing directivity of the T= 0.707 LeCleac'h horn and the 1.4" throat size of the large-format driver. Unlike a small-format horn which aims at reaching 15 kHz or above, this horn gives up the top octave in return for another octave at the bottom of the range. The diaphragm of a large-format driver is twice as large as a small-format driver, and that's the tradeoff.
You wouldn't expect a 15" bass driver to behave in the same way as a 7" midbass; they cover different frequency ranges, and have different power-handling characteristics. Similarly, the large and small-format drivers do different tasks.
I've built speakers with supertweeters before, and the crossover is a nuisance to measure, since the drivers are several wavelengths apart. This makes microphone positioning rather critical; move it up or down a few inches, and the response shows the typical notches of a speaker with tightly spaced vertical lobing. Intuitively, this seems like a bad thing, since you know the listening height is not that well controlled, and the listener might even tilt their head enough to create dissimilar responses at each ear.
But - in practice, it is remarkably difficult to hear. It really is, although very obvious in measurement. This is just like time distortion of square waves; it looks just dreadful on a scope, yet is not easy to hear. I've flipped the switch on all-pass filters and was hard-pressed to hear any difference - at most, a very subtle change in tonality, nothing like the clear change a 1/2 dB level difference in the midrange can make.
The supertweeter crossover, at least if it reasonably competent, is also more subtle than you might expect. The one thing to avoid is having the supertweeter louder than the main speaker; the shelved-up response is subjectively experienced as a peak at the frequency where the shelf begins. So if there's going to be an error, it should always be in the direction of having the supertweeter slightly down in level, and never higher. That does sound unnatural.
This isn't surprising; in nature, there are very few mechanisms that elevate treble over bass, while merely passing sound through 50 feet or more of air will dull the upper treble. A slight dullness merely sounds like the music is further away. Trim the supertweeter level, and you adjust the sense of "air" at the top - go too far, and the sound becomes unnatural and artificial-sounding.
One unusual aspect of ribbon tweeter's narrow vertical directionality is they don't fall off with distance as fast as direct-radiators. Although a ribbon and a dome tweeter might have equal levels at 1 meter, at 3 meters the ribbon pulls ahead by several dB because of less attenuation with distance.
In practice, you never want to equalize a speaker at 1 meter; the only speakers that are listened at that distance are nearfield monitors mounted on top of the meter bridge of a recording console. This system is never going to mounted on top of a console, nor will it be listened to at 1 meter. A 2.5 to 4 meter listening distance should be more typical.
If the directivity index of all the drivers were identical and invariant with frequency, sure, we could equalize at any distance and the speaker would behave just the same. In practice, though, the drivers have different directivity indexes, particularly at the edges of their working ranges. This means the final balancing needs to account for listening distance, or at the least, choose an average distance of say, 3 meters, and balance the speaker for that.
I might as well step in the big mud puddle of the directivity debate, now that I've gone this far. I believe in equalizing the speaker for the most likely listening positions, but do not agree with modern constant-directivity theory.
The reason is fairly simple; the first-arrival sound dominates the perception of spectral balance, while the first reflections come from wildly different emission angles, far outside the usual 60~90 degree "good window" of constant-directivity designs. We don't have to guess where the first reflections are coming from; we know what they are.
The first is the floor bounce; this will arrive 2.5~3 mSec after the direct sound. Most people will use nothing more than carpet (and foam or rubber underlayment) to absorb floor bounce. I've measure the spectra of the reflection off heavy carpet, and there is essentially no absorption below 7 kHz. So there floor bounce is there, whether we want it or not.
If the listener is at 48 inches (unlikely), the center of the speaker is also at 48 inches, and the ceiling is 8 feet, the ceiling reflection will arrive at the same time as the floor bounce. With a lower listening position and a higher ceiling, the ceiling reflection will arrive 1 mSec or more after the floor bounce.
Side walls and rear walls: if the speaker is jammed in a corner, the nearest sidewall reflection will be in the same range as the floor bounce - for obvious and some non-obvious reasons, this is not desirable. The same applies to the rear wall reflection - if this arrives at the same time as the floor bounce, or side wall reflection, this creates a noticeable coloration, since the reflections add up, and deepen the comb-filter nulls and peaks. In practice, moving the speaker out a moderate distance into the room spreads the reflections apart, leaving the floor bounce as the first reflection that arrives at the listening position. The others - ceiling, sidewall, and rear wall - follow several milliseconds later. Multiple bounces arrive much later, 10 mSec or more, unless the room is very small.
All four reflections have one thing in common: the emission angles for these first reflections are well outside the 60-90 degree controlled-directivity radiation angle. In fact, you can't hear the emission of the speaker 10 or 20 degrees away from where you're sitting; you have to physically move there to intercept the sound with your ears. When you sit in one location, you hear the sound that's coming right at you, plus the four reflections mentioned above, then an array of later reflections. All of the reflections take direct paths from the speaker to the reflection surface to you, and the great majority of emission angles for these reflections are far outside the 60-90 degree "good performance" window.
As far as nearly all these reflections are concerned, the emission angle is nearly random - not just from the front of the speaker, but the sides, top, bottom, and rear are equally weighted in terms of room reflections.
That is why I give greatest weight to the first-arrival sound - for my preference, maybe a weighting of 60~75% in terms of design priorities, which is quite different than the modern school of constant-directivity loudspeaker design. But these are the reasons why: you can't sit in five places at once, and domestic loudspeakers are not intended theater or PA use, where audience coverage is the prime directive. Sure, the listener in the living room might be well off-axis, and the dispersion should be smooth and even enough to cover all likely listening positions. But the energy going into the room reflections isn't coming from the preferred 60-90 emission angle; it's coming from all over the speaker, not just the forward radiation pattern.
So I give first priority to direct-arrival sound, and then consider the spectra of the total power of the speaker radiated into a sphere, since that is what is powering the reflections - they really do come from all directions at once, not just a preferred emission angle in the forward direction. With a multiway speaker with non-coincident drivers, the time profile of the speaker will be radically different for every reflection, since the emission angles are so radically different. The time profile that matters is the first arrival; all the rest will be thoroughly scrambled, unless you have a near-perfect single-point emitter (with no crossovers).
Last edited:
Really good post!
I'll add-in a few things..
Going from a 1.75" (1 inch exit) to a 3" (1.4 inch exit) is almost 3 times the area.
Larger diameter compression drivers exhibit less edge damping for their respective surface area then smaller drivers. This tends to improve the subjective sense of decay and field depth.
Dispersion of direct sound is still an issue. We don't hear a single axis, rather we hear (or don't hear as the case may be, particularly in relation to freq.), the full 360 degree pattern. Of course certain ranges are far more dominate, particularly +/- 30 degrees horizontally forward relative to the listener. Surprisingly I remember reading the the next most dominate range is +/- 30 degrees reward. (..in other words we tend to focus on sound horizontally in a bi-pole fashion.)
Thoroughly agree with this. I find that my listening impression corresponds with the on axis response and specifically not with the off axis response.
It's comforting to know, and reflects my own experiences quite well. Once I had a setup of very high-reaching "midrange" which played up to 12~14kHz with a supertweeter. With so large the c-c distance, I didn't notice any negative effect at the time.
One step further, my current system is also suffering from large c-c distance between mid and tweeter and it's not supertweeter above 7kHz, xover'ed @ about 3k instead
Of course the sharpness of image is somewhat inferier compared with a wideranger covering the whole range, but it's not annoying at all (for me, at least). Its strengths in other aspects stand up to this 'little' drawback.However I'm not so sure about the directivty things (or I should say average 'space response' ?). My current (fiddled) system consists of mid and tweeter horns of different profiles. The tweeter is CD WG and the mid is tractrix. I do hear the difference bwtween them, off-axis. CD indeed plays more constantly in larger listening area, while the changes (off axis) in the tractrix is clearly audible.
In my very live living room, flatter space response (by constant directivity) is welcome. Because I hear very much relfection. I let the mid horn open-back'ed and add additional rear-firing tweeter (also in WG). They help quite a lot in this regard. I've tried mid with sealed back and no rear-firing tweeter, but the overall presentation is simply inferior. Even with very near rear wall, I still prefer the exist of rear firing sound - it's more natural as a whole, also more stable in tonal quality - simply put it's more 'complete' and is more like the 'real things in room'. The rear sounds (and reflections all over the place) make the large c-c distance between mid and tweeter merge better and even less distinguishable. Does the rear sound mess up the imaging? I don't feel that way and I don't think so.
Been struggling all day with the audio part of the HT upgrade, replacing the obsolete Denon 2805 with Marantz AV8003 pre/pro and MM8003 eight-channel amplifier. Not that I'm using eight channels; it's a conventional 5.1 setup, with the Ariels providing Left and Right, Dynaudio CS X center, a pair of very mediocre Tannoy M1's for the rears, and a REL Strata II subwoofer. (No good location for surround rears, otherwise I'd be using them.)
No problem with the wiring per se, just the immense weight of the 60" wide Salamander rack with the Panasonic P58VT25 mounted on the rear frame. With the 95 lb. plasma TV, who knows how much for the rack, and the combined 77 lb. weight of the two Marantz's, this sucker must weigh 300 lbs. Yes, it has wheels, but getting them going on thick carpet is a bear.
Since the rack - for WAF reasons - is only 8" out from the wall, the whole shebang has to be wheeled out for access. And I'm not that petite at 6' 2", so it has to wheeled way out into the room to give enough space to access the zillions of wires connecting all this stuff together. Sigh, it would be nice to have a media closet where all the wiring would be oh-so-easy, just like rewiring rack-mount equipment. Open the door and there all the connections are, just waiting for the plug-and-play experience.
After the Big Pull - grunt - let's see what goes where: HDMI 1.4 from the Blu-Ray player to the TV, HDMI 1.3a from the secondary-audio output of the player to the Marantz, HDMI 1.3a from the Comcast DVR to the TV, and HDMI 1.3a from the Marantz to the TV (so I can set it up on the TV screen). Optical S/PDIF from the Comcast DVR to the Marantz, another optical from the TV to the Marantz (so I can listen to Pandora on the hifi system). Ethernet to the TV and Blu-Ray player (firmware updates, doncha know). The Marantz has Ethernet too, but it's only for for DLNA, which doesn't work that great with Macintosh computers (no AIFF support). Five XLRs from the AV8003 to the MM8003, along with a single RCA-to-RCA to synchronize on-off powering of both units. Another single RCA to the subwoofer. Oh yes, the five sets of speaker wires too, bi-wired for the Ariels. Lotsa wires.
The MM8003 has those very annoying EU-compliant speaker jacks which only accept bare wire or banana jacks - how on Earth could anyone ever harm themselves with speaker-level voltages - by sucking on the wires maybe? Also annoying that some of the speaker jacks on the Marantz MM8003 bend a little when I slide in the heavy Blue Jeans Cable locking banana plugs - don't like that. Also got crosswise of the locking banana plugs - not enough room to work, put some in the wrong channel, try and remove them, some get stuck, some spin around and still stay stuck, grrr, this is frustrating! Why O why did the EU condemn millions of people to using with these dreadful banana jacks. Are speakers and power amps really that unsafe? Has anyone in the whole wide world ever been injured or killed by a loudspeaker or power amp connection? Thousands die on the highways all over the world, maybe we should stop driving, too? And bicycles, just think how dangerous they are!
Aside from being tired and not in the best mood (as you can probably tell), all went well until the very slow roll-back, moving only a few inches at a time, sweeping and clearing the wires as I go, when I discovered to my dismay one of the rear wheels crushed a wire I somehow missed. At first I thought it was one of the brand-new HDMI 1.4 wires. Argh! Those aren't cheap, and don't tolerate that kind of abuse. But fortunately it was merely a cheap RCA coax for the subwoofer - I have lots of those, so I replaced it in a jiffy. (Never a good idea to leave an abused wire in-circuit, it could fail at any time, with unpleasant results for the electronics.) Next time I do this I'm going to hang all of these silly wires from some kind of hanger off the rear of the Salamander rack, so they don't lay on the floor, asking for trouble.
Too physically tired, mind-fogged, and cranky to finish the install tonight - the screwup with the wire was a good warning to stop before I did any more damage. Tomorrow, thoroughly double-check everything before powering up, and run the Audyssey auto-install (modern HT receivers insist on running Audyssey or equivalent on initial setup), so I want to be fresh as I step through measuring the speakers. Not all that impressed with Audyssey equalization, but the pre/pro will insist, so I'll give it what it wants. Maybe for movies, or setting it so the other speakers mirror the balance of the Ariels.
Gary managed to divert me with a great phone call about his progress on the HF crossover. No pix to show you guys, but it's looking good, particularly for a large-format compression driver with a 3" aluminum diaphragm. 3rd-order highpass crossover, Dave Slagle autoformer attenuator, and Zobel corrector are working as expected. From what I see, it will be a low-coloration high-efficiency (97~98 dB/meter/watt) system with generous dynamic range.
No problem with the wiring per se, just the immense weight of the 60" wide Salamander rack with the Panasonic P58VT25 mounted on the rear frame. With the 95 lb. plasma TV, who knows how much for the rack, and the combined 77 lb. weight of the two Marantz's, this sucker must weigh 300 lbs. Yes, it has wheels, but getting them going on thick carpet is a bear.
Since the rack - for WAF reasons - is only 8" out from the wall, the whole shebang has to be wheeled out for access. And I'm not that petite at 6' 2", so it has to wheeled way out into the room to give enough space to access the zillions of wires connecting all this stuff together. Sigh, it would be nice to have a media closet where all the wiring would be oh-so-easy, just like rewiring rack-mount equipment. Open the door and there all the connections are, just waiting for the plug-and-play experience.
After the Big Pull - grunt - let's see what goes where: HDMI 1.4 from the Blu-Ray player to the TV, HDMI 1.3a from the secondary-audio output of the player to the Marantz, HDMI 1.3a from the Comcast DVR to the TV, and HDMI 1.3a from the Marantz to the TV (so I can set it up on the TV screen). Optical S/PDIF from the Comcast DVR to the Marantz, another optical from the TV to the Marantz (so I can listen to Pandora on the hifi system). Ethernet to the TV and Blu-Ray player (firmware updates, doncha know). The Marantz has Ethernet too, but it's only for for DLNA, which doesn't work that great with Macintosh computers (no AIFF support). Five XLRs from the AV8003 to the MM8003, along with a single RCA-to-RCA to synchronize on-off powering of both units. Another single RCA to the subwoofer. Oh yes, the five sets of speaker wires too, bi-wired for the Ariels. Lotsa wires.
The MM8003 has those very annoying EU-compliant speaker jacks which only accept bare wire or banana jacks - how on Earth could anyone ever harm themselves with speaker-level voltages - by sucking on the wires maybe? Also annoying that some of the speaker jacks on the Marantz MM8003 bend a little when I slide in the heavy Blue Jeans Cable locking banana plugs - don't like that. Also got crosswise of the locking banana plugs - not enough room to work, put some in the wrong channel, try and remove them, some get stuck, some spin around and still stay stuck, grrr, this is frustrating! Why O why did the EU condemn millions of people to using with these dreadful banana jacks. Are speakers and power amps really that unsafe? Has anyone in the whole wide world ever been injured or killed by a loudspeaker or power amp connection? Thousands die on the highways all over the world, maybe we should stop driving, too? And bicycles, just think how dangerous they are!
Aside from being tired and not in the best mood (as you can probably tell), all went well until the very slow roll-back, moving only a few inches at a time, sweeping and clearing the wires as I go, when I discovered to my dismay one of the rear wheels crushed a wire I somehow missed. At first I thought it was one of the brand-new HDMI 1.4 wires. Argh! Those aren't cheap, and don't tolerate that kind of abuse. But fortunately it was merely a cheap RCA coax for the subwoofer - I have lots of those, so I replaced it in a jiffy. (Never a good idea to leave an abused wire in-circuit, it could fail at any time, with unpleasant results for the electronics.) Next time I do this I'm going to hang all of these silly wires from some kind of hanger off the rear of the Salamander rack, so they don't lay on the floor, asking for trouble.
Too physically tired, mind-fogged, and cranky to finish the install tonight - the screwup with the wire was a good warning to stop before I did any more damage. Tomorrow, thoroughly double-check everything before powering up, and run the Audyssey auto-install (modern HT receivers insist on running Audyssey or equivalent on initial setup), so I want to be fresh as I step through measuring the speakers. Not all that impressed with Audyssey equalization, but the pre/pro will insist, so I'll give it what it wants. Maybe for movies, or setting it so the other speakers mirror the balance of the Ariels.
Gary managed to divert me with a great phone call about his progress on the HF crossover. No pix to show you guys, but it's looking good, particularly for a large-format compression driver with a 3" aluminum diaphragm. 3rd-order highpass crossover, Dave Slagle autoformer attenuator, and Zobel corrector are working as expected. From what I see, it will be a low-coloration high-efficiency (97~98 dB/meter/watt) system with generous dynamic range.
Last edited:
What you think would be the difference in usable extension on the bottom and top end between a 1.4" and a bigger 2" throat? I ask because I've got a AH425 and a 2" Radian 850PB on the way.
I'm debating crossover points to a TD15M bellow and a 465PB supertweeter above and this thread had been very useful - thanks guys.
A) Be glad you have HDMI and don't have to use component Video with separate stereo audio cables.
B) Ever put you finger across a 10,000uf cap charged to 50 volts?
Lynn, whatever happened with the LO15's? I'm working with the Dipole 15's right now and am impressed with the low end output on my small, simple baffle, but am wondering if the LO15 would possibly sound cleaner on the low end due to lower QTS. Anything you can report?
Edit: My mistake, looking back to the specs, they seem to share the same QTS, so there goes that line of reasoning. Still curious though Lynn.
See here for the current prototype:
Lambda Dipole 15" measurements - Techtalk at Parts-Express.com
Greg Begland - technician - Phat Planet Recording Studios - Orlando Florida - Audiophile Quality Recording, Mixing, Mastering, Production
Edit: My mistake, looking back to the specs, they seem to share the same QTS, so there goes that line of reasoning. Still curious though Lynn.
See here for the current prototype:
Lambda Dipole 15" measurements - Techtalk at Parts-Express.com
Greg Begland - technician - Phat Planet Recording Studios - Orlando Florida - Audiophile Quality Recording, Mixing, Mastering, Production
Last edited:
Still some VIFA on the market ?
Hi All,
I am following this post since years but still no successor to the Ariel ;-)
I built my Ariels 2 years ago and now my father wants the same (guess why).
Unfortunately the VIFA P13WH-00-08 are no more produced nor sold.
That's why I am asking to the community if someone have 4 of them available ?
Lynn I send you a PM asking for an alternative but I guess there is no chance to have a perfect match with another reference.
So if somebody can help, feel free to PM me.
Thanks
Hi All,
I am following this post since years but still no successor to the Ariel ;-)
I built my Ariels 2 years ago and now my father wants the same (guess why).
Unfortunately the VIFA P13WH-00-08 are no more produced nor sold.
That's why I am asking to the community if someone have 4 of them available ?
Lynn I send you a PM asking for an alternative but I guess there is no chance to have a perfect match with another reference.
So if somebody can help, feel free to PM me.
Thanks
First Question: I've sold my first pair of LO15's, and will sell the second pair as well. I custom-ordered four of them, and the project has taken a turn away from the OB concept due to my unwillingness to go down the path of active equalization.
The change in direction of the project is not a reflection on the drivers; from the feedback I've gotten, the drivers are sonically outstanding, with the hallmark TD15M sound in an OB driver. Many people like OB's, and I'm pleased that AESpeakers is offering these drivers as part of their product line. The attitude I have against equalization is my problem; many people are happy with digital or analog equalized OB systems.
Second question: The Ariel was a good design for its day, and I waited patiently for Scan-Speak/Vifa/Seas drivers that were more efficient and just as flat as the 5.5" Vifas that were an essential part of the Ariels. Well, that day never came.
The drivers that followed were no more efficient, and were less flat thanks to the fad of more rigid but also more resonant cone materials - which in turn demanded much more complex crossovers to remove the artifacts of the breakup region. The key concept in the Ariel - which is applied in the new project - are wideband drivers with well-controlled rolloff regions, thus simplifying the crossover and also allowing a higher level of parts quality. Bass and midbass drivers with well-controlled rolloff regions are still pretty rare, and more rare in the high-efficiency realm.
As for drivers that are closest in spirit to the 5.5" Vifas, and at a higher level of quality, I'd look at the Skaaning Flex-Units. I've heard these, and they are very, very good, although not as cheap as the original Vifas. But Skaaning has the right idea; drivers with very good self-damping that don't require elaborate crossovers.
The initial listening session of the Marantz AV8003 and MM8003 was underwhelming; plenty of power to be sure, and smooth with little evidence of transistor grain, but rather smudged and constrained-sounding. At four to five times the price of the Denon 2805, I was having mixed feelings about the purchase. But Karna and I kept listening in the small hours of the morning, and it very gradually started sounding better - more dynamic, more open, and quicker-sounding.
The next day, when twiddling around with the assorted surround settings, I noticed that even at very quiet levels, it was sounding much clearer and more open than the previous night. Karna and I listened at more length the following night, and the sound was much better all around. The most noticeable quality is the way the soundfield extends well into the room - with a lot of surround gear, the sound is pasted against the wall, with very little sense of depth. The subtleties of timbre are pretty good for solid-state gear, and there's oodles of power at a measured 170 watts per channel. This is more to my taste than the Anthem gear, which I didn't warm up to.
So I guess there's something to break-in after all. My only guess is the dielectrics - in the new cabling and speaker wire, the circuit boards, and the electrolytics in the power supply and coupling caps (in the preamp) - are subtly altering with applied voltage over time. The units do emit that distinctive Chinese circuit board smell, so something is happening to the passivation layer of the circuit board.
The Audyssey auto-equalization was a big flop; I'll be trying it again, but measuring from three different points along the front of the couch produced results much inferior to the five-year-old Denon with its single-point equalization. Auditioning with surround material that used smooth circular pan revealed the rear speakers were too hot by 3 to 4 dB, a terrible result, and it attempted to add a shocking 5 dB or more of treble boost to rear speakers that were already thin and tinny-sounding.
I was gratified that the Dynaudio CS X center speaker measured within a dB or so of the Ariels, making EQ across the front largely unnecessary. More important, it sounded much better with the EQ off, and more surprisingly, with distance compensation for the center speaker turned off as well. So aside from distance-compensating the rears (which needed it), I bypassed the equalizer.
I suspect the equalization is not smart enough to measure the first-arrival sound by itself, and is simply doing a crude overall room correction, lumping the first-arrival and later sounds together, like a real-time analyzer. It also tries to "fill in" nulls, which I don't appreciate. Left to its own devices, it produces far too much rear energy - quite audibly so on front-to-rear pans - and wants to crank up the treble of the whole system, converting the flat response of the Ariels to the thin, bright sound of modern theaters. No thanks.
Maybe the latest-and-greatest Audyssey system is better; I don't know. But the underlying assumption of measuring a zillion different points seems wrong in principle, and seems to strongly imply the system is indeed aimed at room correction, and takes no account at all of the spectral response of the first-arrival signal. So you can start with speakers that are quite flat and end up with massive boosting after the auto-equalizer gets through with it.
With ordinary speakers that have increasing directivity at higher frequencies, if you attempt to flatten out the overall room energy, the first-arrival (on-axis) sound will be heavily overboosted at high frequencies, and will sound shrill and bright. Some reviewers seem to have the impression that this kind of sound is "accurate" - well, maybe if your standard of comparison is a PA system. But acoustic music, sans amplification, is very rarely shrill and bright - it sounds shockingly dull compared to the sound of the typical hifi show.
My ideal version of an equalizer would confine its room equalization to the region below 300 Hz, would only equalize first-arrival sounds above that frequency, and would not attempt to equalize narrowband nulls, but simply gracefully ignore them. Even nicer would be FIR time correction for the above-300 Hz region, allowing passive crossovers to be "unwrapped" in the phase domain. But from what I see in the HT arena, the auto-equalizers are nowhere close to this degree of sophistication, and are probably best avoided in systems that have speakers with flat responses.
But then again, looking at the appalling measurements of expensive high-end speakers reviewed in $tereophile magazine, maybe these gizmos are needed, at least for the HT market, where many Center and surround speakers have grossly nonflat responses.
The change in direction of the project is not a reflection on the drivers; from the feedback I've gotten, the drivers are sonically outstanding, with the hallmark TD15M sound in an OB driver. Many people like OB's, and I'm pleased that AESpeakers is offering these drivers as part of their product line. The attitude I have against equalization is my problem; many people are happy with digital or analog equalized OB systems.
Second question: The Ariel was a good design for its day, and I waited patiently for Scan-Speak/Vifa/Seas drivers that were more efficient and just as flat as the 5.5" Vifas that were an essential part of the Ariels. Well, that day never came.
The drivers that followed were no more efficient, and were less flat thanks to the fad of more rigid but also more resonant cone materials - which in turn demanded much more complex crossovers to remove the artifacts of the breakup region. The key concept in the Ariel - which is applied in the new project - are wideband drivers with well-controlled rolloff regions, thus simplifying the crossover and also allowing a higher level of parts quality. Bass and midbass drivers with well-controlled rolloff regions are still pretty rare, and more rare in the high-efficiency realm.
As for drivers that are closest in spirit to the 5.5" Vifas, and at a higher level of quality, I'd look at the Skaaning Flex-Units. I've heard these, and they are very, very good, although not as cheap as the original Vifas. But Skaaning has the right idea; drivers with very good self-damping that don't require elaborate crossovers.
The initial listening session of the Marantz AV8003 and MM8003 was underwhelming; plenty of power to be sure, and smooth with little evidence of transistor grain, but rather smudged and constrained-sounding. At four to five times the price of the Denon 2805, I was having mixed feelings about the purchase. But Karna and I kept listening in the small hours of the morning, and it very gradually started sounding better - more dynamic, more open, and quicker-sounding.
The next day, when twiddling around with the assorted surround settings, I noticed that even at very quiet levels, it was sounding much clearer and more open than the previous night. Karna and I listened at more length the following night, and the sound was much better all around. The most noticeable quality is the way the soundfield extends well into the room - with a lot of surround gear, the sound is pasted against the wall, with very little sense of depth. The subtleties of timbre are pretty good for solid-state gear, and there's oodles of power at a measured 170 watts per channel. This is more to my taste than the Anthem gear, which I didn't warm up to.
So I guess there's something to break-in after all. My only guess is the dielectrics - in the new cabling and speaker wire, the circuit boards, and the electrolytics in the power supply and coupling caps (in the preamp) - are subtly altering with applied voltage over time. The units do emit that distinctive Chinese circuit board smell, so something is happening to the passivation layer of the circuit board.
The Audyssey auto-equalization was a big flop; I'll be trying it again, but measuring from three different points along the front of the couch produced results much inferior to the five-year-old Denon with its single-point equalization. Auditioning with surround material that used smooth circular pan revealed the rear speakers were too hot by 3 to 4 dB, a terrible result, and it attempted to add a shocking 5 dB or more of treble boost to rear speakers that were already thin and tinny-sounding.
I was gratified that the Dynaudio CS X center speaker measured within a dB or so of the Ariels, making EQ across the front largely unnecessary. More important, it sounded much better with the EQ off, and more surprisingly, with distance compensation for the center speaker turned off as well. So aside from distance-compensating the rears (which needed it), I bypassed the equalizer.
I suspect the equalization is not smart enough to measure the first-arrival sound by itself, and is simply doing a crude overall room correction, lumping the first-arrival and later sounds together, like a real-time analyzer. It also tries to "fill in" nulls, which I don't appreciate. Left to its own devices, it produces far too much rear energy - quite audibly so on front-to-rear pans - and wants to crank up the treble of the whole system, converting the flat response of the Ariels to the thin, bright sound of modern theaters. No thanks.
Maybe the latest-and-greatest Audyssey system is better; I don't know. But the underlying assumption of measuring a zillion different points seems wrong in principle, and seems to strongly imply the system is indeed aimed at room correction, and takes no account at all of the spectral response of the first-arrival signal. So you can start with speakers that are quite flat and end up with massive boosting after the auto-equalizer gets through with it.
With ordinary speakers that have increasing directivity at higher frequencies, if you attempt to flatten out the overall room energy, the first-arrival (on-axis) sound will be heavily overboosted at high frequencies, and will sound shrill and bright. Some reviewers seem to have the impression that this kind of sound is "accurate" - well, maybe if your standard of comparison is a PA system. But acoustic music, sans amplification, is very rarely shrill and bright - it sounds shockingly dull compared to the sound of the typical hifi show.
My ideal version of an equalizer would confine its room equalization to the region below 300 Hz, would only equalize first-arrival sounds above that frequency, and would not attempt to equalize narrowband nulls, but simply gracefully ignore them. Even nicer would be FIR time correction for the above-300 Hz region, allowing passive crossovers to be "unwrapped" in the phase domain. But from what I see in the HT arena, the auto-equalizers are nowhere close to this degree of sophistication, and are probably best avoided in systems that have speakers with flat responses.
But then again, looking at the appalling measurements of expensive high-end speakers reviewed in $tereophile magazine, maybe these gizmos are needed, at least for the HT market, where many Center and surround speakers have grossly nonflat responses.
Last edited:
As for HDMI versus the prior generation of Component-color cables, yes, I've paid my dues there. The previous Proscan/RCA/Thomson multisync 36" TV only had one set of component inputs, so I had to mess around with an awkward component switcher to accommodate the various HD sources, not to mention the five-cable bundles. No SCART connections in NTSC-land. It was a long march from composite/RCA to S-Video to Component to HDMI.
HDMI is kind of a mixed blessing. The HDCP authentication procedure takes a long time - several seconds, so switching between sources involves what seems like a very long blackout period on the television. Karna has asked if something is wrong with the TV when I switch sources. I know the HDCP authentication procedure involves a lot of handshaking between devices to prevent us evil consumers from pirating Michael Bay's little art productions, but why does it takes so long with processors operating in the GHz region and cabling capable of gigabit speeds? Where's the holdup, guys? Is there a software process that's timing-out somewhere? Why can't authentication happen in one video frame? This seems like mighty poor design.
I do like that we finally have an industry-standard connection and protocol that supports high-resolution digital audio as well as DSD. It could have been Firewire/IEEE1394, but that was not to be. During the long wait, DVD-A and SACD gradually weakened, and it's an open question if audio-only Blu-ray discs will make it in a marketplace dominated by ever-lower quality formats. Still, for all its faults, HDMI is finally here.
HDMI is kind of a mixed blessing. The HDCP authentication procedure takes a long time - several seconds, so switching between sources involves what seems like a very long blackout period on the television. Karna has asked if something is wrong with the TV when I switch sources. I know the HDCP authentication procedure involves a lot of handshaking between devices to prevent us evil consumers from pirating Michael Bay's little art productions, but why does it takes so long with processors operating in the GHz region and cabling capable of gigabit speeds? Where's the holdup, guys? Is there a software process that's timing-out somewhere? Why can't authentication happen in one video frame? This seems like mighty poor design.
I do like that we finally have an industry-standard connection and protocol that supports high-resolution digital audio as well as DSD. It could have been Firewire/IEEE1394, but that was not to be. During the long wait, DVD-A and SACD gradually weakened, and it's an open question if audio-only Blu-ray discs will make it in a marketplace dominated by ever-lower quality formats. Still, for all its faults, HDMI is finally here.
Last edited:
I found this interesting: Haigner Alpha.
haigner.com
http://home.tele2.at/haigner/download/haigner_alpha_horn.pdf
I think we've seen similar mid-treble horns before, but the mid-bass horn is a bit different.
While it is rather large, it's also quite aesthetically pleasing IMO.
Here is another mid-bass horn from Haigner (gamma) (..3 pic.s, the last pic on the page is one of them with their soon-to-be owner):
The Audio Eagle - The Vienna Vibes Audio Fest
haigner.com
http://home.tele2.at/haigner/download/haigner_alpha_horn.pdf
I think we've seen similar mid-treble horns before, but the mid-bass horn is a bit different.
While it is rather large, it's also quite aesthetically pleasing IMO.
Here is another mid-bass horn from Haigner (gamma) (..3 pic.s, the last pic on the page is one of them with their soon-to-be owner):
The Audio Eagle - The Vienna Vibes Audio Fest
Attachments
Last edited:
